[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-18 Thread bobv
Patch updated. http://gwt-code-reviews.appspot.com/473801/diff/78001/79002 File dev/core/src/com/google/gwt/dev/shell/CompilingClassLoader.java (right): http://gwt-code-reviews.appspot.com/473801/diff/78001/79002#newcode695 dev/core/src/com/google/gwt/dev/shell/CompilingClassLoader.java:695: *

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-18 Thread scottb
LGTM http://gwt-code-reviews.appspot.com/473801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-17 Thread Scott Blum
Bob, I'm looking at this in depth right now, but I did have a high level question: On Thu, Jun 17, 2010 at 12:02 AM, b...@google.com wrote: - Methods with JsoSubtype[] parameters have their names mangled - Constructors have a synthetic disambiguator parameter added since we can't change

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-17 Thread bobv
Just curious, why not simply use the same technique in both cases? The disambiguator parameter should work find for non-constructor methods, right? Adding the extra parameter means fiddling with the stack to push the extra null argument. Since non-constructor method invocation occurs more

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-17 Thread Scott Blum
On Thu, Jun 17, 2010 at 11:15 AM, b...@google.com wrote: Just curious, why not simply use the same technique in both cases? The disambiguator parameter should work find for non-constructor methods, right? Adding the extra parameter means fiddling with the stack to push the extra null

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-17 Thread Scott Blum
On Thu, Jun 17, 2010 at 11:24 AM, Scott Blum sco...@google.com wrote: Tough call. :( I'll trust your judgement that the speed is worth the extra complexity. On the other hand looking through this patch and stepping through the debugger, it looks like the extra overhead of pushing a few

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-17 Thread scottb
Mostly LG, just some nits and food-for-thought. http://gwt-code-reviews.appspot.com/473801/diff/78001/79002 File dev/core/src/com/google/gwt/dev/shell/CompilingClassLoader.java (right): http://gwt-code-reviews.appspot.com/473801/diff/78001/79002#newcode695

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-06-16 Thread bobv
Another iteration of the patch that simplifies the cross-casting of JSO subtype arrays by upcasting all JSO subtype arrays to JavaScriptObject[]. This ensures that array object identity will be preserved, even after an illegal cast between incompatible array types. Changes to look at: - Added

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-05-26 Thread bobv
Ready for another look. By adding an actual @SingleJsoImpl annotation to act as a hint for devmode, the impact on I18N's CurrencyData is pretty much a one-liner. http://gwt-code-reviews.appspot.com/473801/diff/5001/6004 File dev/core/src/com/google/gwt/dev/shell/JsValueGlue.java (right):

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-05-21 Thread scottb
My brain hurts now. http://gwt-code-reviews.appspot.com/473801/diff/5001/6004 File dev/core/src/com/google/gwt/dev/shell/JsValueGlue.java (right): http://gwt-code-reviews.appspot.com/473801/diff/5001/6004#newcode63 dev/core/src/com/google/gwt/dev/shell/JsValueGlue.java:63: if (desiredType !=

[gwt-contrib] Re: RR : Reimplement DevMode JavaScriptObject support (issue473801)

2010-05-12 Thread scottb
Cursory review of the non-dev code. Having .cast() not work would bother me. Will get to the dev code soon. http://gwt-code-reviews.appspot.com/473801/diff/5001/6015 File user/src/com/google/gwt/user/client/DOM.java (right):